The Holocene21(6) 939–950© The Author(s) 2011Reprints and permission:sagepub.co.uk/journalsPermissions.navDOI: 10.1177/0959683611400218hol.sagepub.com

Research paper

Pollen-inferred climate changes andvertical shifts of alpine vegetation beltson the northern slope of the Nyainqentanglha Mountains (centralTibetan Plateau) since 8.4 kyr BP

Quan Li,1 Houyuan Lu,1 Liping Zhu,2 Naiqin Wu,1 Junbo Wang2

and Xinmiao Lu2

AbstractFossil pollen from Nam Co and modern pollen from altitudinal vegetation belts around the lake are investigated to reveal alpine vegetation succession in response to climate changes during the Holocene in the central Tibetan Plateau. The discriminant analysis on 37 topsoil samples shows that pollen samples from alpine steppe at lower elevations (<4800 m) and alpine meadow on upper slopes (4800–5200 m) can be distinguished by their pollen assemblages. Samples from alpine steppe contain more Artemisia (25.1%) and Poaceae pollen (11.5%), whereas those from alpine meadow are dominated by Cyperaceae pollen (>60%). Our result indicates that the pollen ratio of Artemisia to Cyperaceae (A/Cy) can be used as an indicator of the vertical shift of vegetation belts and temperature changes in the central Tibetan Plateau as suggested by previous studies. A history of the vertical shift of vegetation belts on the northern slope of Nyainqentanglha Mountains and climate changes since 8.4 kyr BP are thus recovered by 198 fossil pollen assemblages from a 332 cm core of Nam Co. Paleovegetation reconstructed from fossil pollen assemblages through discriminant analysis shows a general downward shift of altitudinal vegetation belts, suggesting a decline in the temperature trend since 8.4 kyr BP. This result is consistent with the reduction of A/Cy ratios. The fossil pollen record also reveals warm and wet climate during the early to mid Holocene, and cold and dry conditions during the late Holocene in the Nam Co area. A comparison of Holocene climatic reconstructions across the Plateau indicates that termination of maximum moisture at around 6–5.5 kyr BP in our record is associated with the southeastward retreat of the Southwest Monsoon.

Keywordsdiscriminant analysis, Holocene, Nam Co, pollen analysis, Tibetan Plateau, vegetation belt

Introduction

Vegetation distribution in the Tibetan Plateau (TP) with an average

elevation above 4000 m is very vulnerable to climate change (Shen

et al., 2008a; Song et al., 2004; Zheng, 1996). As major archives of

Holocene paleoenvironmental changes, lake sediments in the TP

have been widely studied (e.g. Gasse et al., 1991; Morrill et al.,

2006; Shen J et al., 2005; Wu et al., 2006; Zhu et al., 2009). Among

them, lacustrine pollen records have provided paleovegetation and

paleoclimate reconstructions qualitatively (e.g. Lu et al., 2001;

Shen et al., 2005; Shen J et al., 2005; Sun et al., 1993; Van Campo

and Gasse, 1993; Van Campo et al., 1996), semi-quantitatively

(e.g. Herzschuh et al., 2006a, b; Shen and Tang, 1996; Tang et al.,

2009), and quantitatively (e.g. Herzschuh et al., 2009a, b, 2010a,

b; Shen, 2003; Shen et al., 2006, 2008a, b; Tang et al., 2004). How-

ever, climatic conditions reconstructed from pollen records

revealed the occurrence of regional differences in the TP. For

instance, a descending trend in temperature and an increasing

trend in precipitation occurred from the early to late Holocene in

the northeast (Herzschuh et al., 2009b), whereas warm/wet climate

during the early to mid Holocene and cold/dry during the late

Holocene occurred in the central, southeast, and south Plateau

(e.g. Herzschuh et al., 2006a; Tang et al., 1999; Zhu et al., 2009).

Therefore, more well-dated and high-resolution pollen records are

essential for characterising the nature and causes of spatial patterns

of Holocene climate in the TP. To date, most of the published

Holocene pollen records are located in the margins of the Plateau

(e.g. Gasse et al., 1991; Herzschuh et al., 2006b, b; Kramer et al.,

2010; Shen J et al., 2005; Tang et al., 2004; Van Campo et al.,

1996), only a few are from the central Plateau (e.g. Herzschuh

et al., 2006a; Shen et al., 2008a; Sun et al., 1993).

400218 HOL21610.1177/0959683611400218Li et al.The Holocene

1Key Laboratory of Cenozoic Geology and Environment, Institute of Geology and Geophysics, Chinese Academy of Sciences, China

2Key Laboratory of Tibetan Environment Changes and Land Surface Processes, Institute of Tibetan Research, Chinese Academy of Sciences, China

Received 11 August 2010; revised manuscript accepted 12 November 2010

Corresponding author:Quan Li, Key Laboratory of Cenozoic Geology and Environment, Institute of Geology and Geophysics, Beijing 100029, China.Email: quanli_cas@163.com

940 The Holocene 21(6)

The ecotone of meadow and steppe in the central TP is sensi-

tive to climate change (Shen et al., 2008a). Shen et al. (2008a)

revealed a history of the horizontal shift of this ecotone in response

to the intensity of the Southwest monsoon during the mid–late

Holocene based on a pollen record from a large lake (Co Ngion) in

the central TP. A pollen record from another large lake (Zigetang)

in the central Plateau revealed a major vegetation change from

‘warm temperate steppe’ (Artemisia-dominated) to alpine steppe

(Cyperaceae-dominated) in the mid Holocene (Herzschuh et al.,

2006a). They attributed this change to a drop in temperature. Both

studies focused on the horizontal movement of vegetation com-

munities. Up to now, little attention has been paid to the altitudinal

vegetation zones and their variations in response to climate

changes, although they develop well in the mountainous regions of

the TP and its peripheries under fragile high-alpine environmental

conditions (Herzschuh et al., 2006b; Miehe and Miehe, 2000).

The Nam Co is located in the meadow-steppe transitional

region of TP, lying on the northern foot of Nyainqentanglha

Mountains occupied by altitudinal vegetation belts consisting of

non-forest alpine vegetation types. Previous paleoclimate studies

in Nam Co and its adjacent areas mainly focused on long-term

environmental changes based on lake terrace deposits (e.g. Wu

et al., 2004; Zhao et al., 2003; Zhu et al., 2004), or lacustrine sedi-

ments (e.g. Li et al., 2008; Lin et al., 2008; Mügler et al., 2010;

Xie et al., 2009; Zhu et al., 2008). Herrmann et al. (2010) anal-

ysed fossil pollen and other palynomorphs from a set of peat

samples to reconstruct the Holocene vegetation changes in the

Nam Co area. However, their results are limited in explanation of

vegetation changes on the regional scale and in time resolution

because of the small spatial scale of the peat pollen source (rela-

tively local) and a small amount of peat samples. To date, there is

still little knowledge of regional vegetation history in the Nam Co

area and its vicinity during the Holocene. In this study, we con-

duct an investigation on modern pollen assemblages from alpine

steppe and meadow around Nam Co and on the northern slope of

Nyainqentanglha Mountains, and present a Holocene lacustrine

pollen record from a deep water core in Nam Co to study shifts of

altitudinal vegetation belts and climate changes in the central TP.

Study areaNam Co (90°15 –91°10 E, 30°40 –31°00 N, 4718 m a.s.l.), the

second largest lake of the TP, is located on northern slope of Nyain-

qentanglha Mountains. It covers 1982 km2 water area and its maxi-

mum depth is about 95 m in the central lake (Figure 1a, b) (Wang

et al., 2009). It has a catchment area of 10 610 km2. It is a closed-

basin lake with no outflow rivers. Precipitation and inflow of glacial

meltwater are the main sources of lake water while evaporation is

its main loss. The mean annual temperature is around 0°C, and

mean annual precipitation about 280 mm in the Nam Co area (Nam

Co Monitoring and Research Station, 90°59 E, 30°46 N, 4730 m

a.s.l.). In this area, westerly winds dominate from January to May

while southerly winds prevail between June and October, showing

the alternate influences of the Westerlies and Southwest Monsoon

on its climate conditions (You et al., 2007).

Alpine steppe and alpine meadow are two major vegetation

types in the central TP (Chen et al., 2011; Institute of Geography,

1990; Lu, 2008; Miehe and Miehe, 2000; Tibetan Investigation

Group, 1988; Zhang et al., 1981). Details of an altitudinal vegetation

zonation in the Nam Co area are described as follows (Figure 1c):

(1) Alpine steppe (<4800 m elevation) occupies the lakeshore

around Nam Co and most of its catchment area. It mainly

consists of Artemisia stracheyi, A. wellbyi, Stipa pur-purea, S. subsessifliflora var. basiplumosa, Carex moor-croftii and Poa, accompanied by Astragalus, Oxytropis,

Festuca, etc.

(2) Alpine meadow (4800–5200 m) forms a dense vegetation

belt with a vertical extent of 400 m on upper slopes of

the Nyainqentanglha Mountains, and occurs in patches

on hills around the lake. It is dominated by Kobresia

sedges (e.g. K. pygmaea, K. humilis) along with Festuca,

Poa, Polygonum macrophyllum, P. viviparum, Caltha sca-posa, etc.

(3) Alpine sparse vegetation (5200–5700 or 5900 m) occurs

below the ice and snow belt. It is mainly composed of

Saussurea, Waldheimia glabra, with some Meconop-sis, Soroseris glomerata, Draba, Rhodiola, Saxifraga,

Thylacospermum, and cushion plants such as Androsace, Arenaria, etc.

In addition, marsh meadow with hummocks of dominant Kobre-sia tibetica (=K. schoenoides, see Zhang, 2004) and Blysmus is

distributed along riverbanks and on low wetland around the lake.

Associated species include K. humilis, Festuca, Polygonum,

Gentiana, and hydrophyte taxa such as Antenoron and Batra-chium bungei.

Material and methodsSampling and datingModern pollen samples (top 2 cm of soil) were collected from

different vegetation types around Nam Co in 2006, including

14 samples from alpine steppe, 7 from alpine meadow, and 16

from marsh meadow. According to the vegetation map (Insti-

tute of Geography, 1990) and field investigation (Lu, 2008),

these topsoil samples cover the major vegetation types in the

Nam Co catchment (Figure 1c). The elevations of alpine

meadow samples are higher than those of alpine steppe by

50–100 m.

Lacustrine core NMLC-1 (59.5 m water depth, core length

332 cm) was obtained in the east part of the lake in May 2005

using a piston corer. An age–depth model for core NMLC-1 has

been developed by Zhu et al. (2008). Twelve samples were used

for AMS 14C dating at the AMS & Quaternary Chronology Labo-

ratory of Peking University. The 14C dates were calibrated to cal-

endar ages using the IntCal04 calibration data (Reimer et al.,

2004) and the OxCal3.1 program (Bronk Ramsey, 2005, https://

c14.arch.ox.ac.uk/embed.php? File=oxcal.html #program, ver-

sion 78, April 2008). 210Pb/137Cs dating and the calculation of

sedimentation rate of a short gravity core (NMCS-10) collected

from the same location were used to estimate the carbon reser-

voir effects. The depth–age sequence of NMLC-1 core was cor-

rected by subtracting this reservoir effect. One date at 190 cm

suggested an age reversal was not used in the final age model.

The dating results show that core NMLC-1 covers the last 8330

years (Zhu et al., 2008).

Sediments for pollen analysis were sampled at 1 cm intervals

at depths of 332–260 cm (time resolution of ~100 yr/sample),

and at 2 cm intervals for the upper parts (~20 yr/sample). In

total, 198 sediment samples were analysed palynologically in

this study.

Li et al. 941

Pollen analysis and numerical methods

Samples were treated with a modified acetolysis procedure (Fægri

and Iversen, 1989), including 10% HCl, 5% KOH, 40% HF, ace-

tolysis treatments, and sieving with a 7μm screen to remove clay-

sized particles. One tablet of Lycopodium marker spores (27 637

grains/tablet) was added to each sample for estimating pollen

concentration. The average number of counted pollen and spores

is 316 grains per sample. Pollen identification was carried out by

comparisons with published literature (Wang et al., 1997; Xi and

Ning, 1994) and our modern pollen collections. Pollen percent-

ages were computed with the sum of all pollen and spores. Pollen

concentrations (grains/g of dried sample), pollen ratio of Arte-misia to Cyperaceae (A/Cy), and sum of percentages of dryness

indicators, including Chenopodiaceae, Cruciferae, Ephedra,Nitraria and Tamaricaceae (Herzschuh et al., 2006b; Tang et al.,

2009; Yu et al., 2001), were calculated for the interpretation of

fossil pollen records.

Figure 1. (a)Vegetation zones of Tibetan Plateau; (b) topographic map and sampling sites in Nam Co area; (c) vegetation map of Nam Co area, and locations of topsoil samples and core NMLC-1. Solid circles in (a) indicate sites of paleoclimate records mentioned in the text. See Table 1 for site information and references. The terrain elevation data set used for (b) is provided by International Scientific & Technical Data Mirror Site, Computer Network Information Center, Chinese Academy of Sciences (http://datamirror.csdb.cn)

942 The Holocene 21(6)

Discriminant analysis has been used as an effective mathemati-

cal technique to reconstruct paleovegetation (Liu and Lam, 1985;

Shen et al., 2005, 2008b; Sugden and Meadows, 1989), and to

investigate the relationship between modern pollen assemblages

and vegetation types on a regional scale (Li et al., 2009; Lynch,

1996). Here, we use discriminant analysis to determine whether

modern pollen assemblages could represent source vegetation

types from an altitudinal vegetation zone on a local scale. The pro-

cedure is as follows. Modern pollen samples are classified into

three a priori groups (alpine steppe, alpine meadow and marsh

meadow) based on vegetation types at sampling sites. Then per-

centage data of major pollen types are applied to establish the dis-

criminant functions and classify samples into predicted groups. The

comparison between a priori groups and predicted groups is used to

assess the consistency of modern pollen assemblages and vegeta-

tion types. Then, the discriminant functions derived from modern

samples are applied to fossil pollen data from core NMLC-1 to pre-

dict the group membership (i.e. vegetation type with the largest

contribution to pollen assemblage) of each fossil sample. Two indi-

ces are used to summarize the results of discriminant analysis. One

is the ‘probability of modern analog’, which represents the similar-

ity between modern or fossil pollen assemblages and predicted veg-

etation types. Another is ‘vegetation zonal index’, which is

converted from probabilities of group membership between the

predicted and the second most probable group (Liu and Lam,

1985). The detailed process and interpretation of results follow the

procedures of Liu and Lam (1985) and Shen et al. (2008b). Soft-

ware used for discriminant analysis is SPSS 10.0 (SPSS Inc., 1999).

ResultsModern pollen assemblagesModern pollen assemblages from various vegetation types show

marked differences (Figure 2). Pollen assemblages from alpine

steppe mainly consist of Artemisia (up to 66.5%, mean 25.1%)

and Poaceae (up to 29.8%, mean 11.5%). Fabaceae (mean 3.3%),

Thalictrum (2.8%), Chenopodiaceae (2.7%), Polygonum (1.8%),

Caryophyllaceae (1.7%) and arboreal pollen types (10.7%) are

commonly present. Percentages of Cyperaceae (13.6%) and total

pollen concentrations (2452 grains/g) of samples from alpine

steppe are much lower than those from meadows.

Pollen assemblages of meadow samples (alpine meadow

and marsh meadow) are characterized by high percentages of

Cyperaceae (up to 95.1%, mean 63.1%). Other pollen types

present include Poaceae (7.6%), Ranunculaceae (>2%), Polyg-onum (1.9%), and Thalictrum (1.9%). Compared with alpine

steppe samples, pollen assemblages of meadows have much

less Artemisia (7.0%). Samples from alpine meadow have the

highest pollen concentrations (7681 grains/g), while marsh

meadow samples have relatively low pollen concentrations

(3432 grains/g).

Pollen ratios of A/Cy are characteristic for individual veg-

etation type. A/Cy values are generally higher than 1.0 in

alpine steppe (mean 4.9), while lower than 1.0 in meadows

(mean 0.1).

Discriminant analysis on modern pollen assemblages

Discriminant analysis of topsoil samples is conducted on per-

centage data of 16 major pollen types (more than 2% at least in

one sample and of ecological importance), including Artemisia,

Poaceae, Cyperaceae, Thalictrum, Polygonaceae, Chenopodia-

ceae, Saussurea-type, Aster-type, Ranunculaceae, Caryophylla-

ceae, Cruciferae, Rosaceae, Fabaceae, Gentiana, Pinus and

Alnus. Three major vegetation types in the Nam Co area are

assigned certain values as their vegetation zonal indices in

advance (value of 1 for alpine steppe, 2 for alpine meadow and 3

for marsh meadow).

Table 1. List of mentioned paleoclimate records in Tibetan Plateau

Code Site name Altitude (m) Latitude and longitude(N, E) Proxies Reference

1 Sumxi Co 5058 34°30 , 81°30 pollen, diatom Van Campo and Gasse (1993)2 Bangong Co 4241 33°40 , 79°00 pollen Van Campo et al. (1996)3 Guliya 6710 35°17 , 81°29 e.g. 18O, dust, Cl– Thompson et al. (1997)4 Selin Co 4530 31°34 -57 , 88°31 -89°21 pollen Sun et al. (1993)5 Zigetang 4560 32°00 , 90°54 pollen Herzschuh et al. (2006a)6 Co Ngion 4515 31°28 , 91°30 pollen Shen et al. (2008a)7 Cuoe Lake 4532 31°30 , 91°32 Sr, TOC, C/N, 13C Wu et al. (2006)8 Ahung Co 4575 31°37 , 92°04 e.g. CaCO3, C/N, 13C Morrill et al. (2006)9 Chen Co 4420 28°53 -59 , 90°33 -39 e.g. grain size, TOC, TN, Fe/

Mn, pollenZhu et al. (2009)

10 Hidden Lake 4980 29°49 , 92°32 pollen Tang et al. (1999)11 Ren Co 4450 30°44 , 96°41 pollen Tang et al. (1999)12 Kusai Lake 4475 35°37 -50 , 93°15 -38 TOC, grain size Liu et al. (2009)13 Dunde 5325 38°06 , 96°24 pollen Liu et al. (1998)14 Qinghai Lake 3200 36°15 -37°32 , 99°36 -100°47 pollen, CaCO3, TOC, TN, 13C Shen J et al. (2005)15 Luanhaizi 3200 37°36 , 101°21 pollen Herzschuh et al. (2006b)16 Koucha Lake 4540 34°00 , 97°12 pollen Herzschuh et al. (2009b)17 Ximencuo 4000 33°23 , 101°06 e.g. TOC, CaCO3, grain-size,

magnetic susceptibilityZhang and Mischke (2009)

18 Hongyuan peat 3466 32°46 , 102°30 13C Hong et al. (2003)19 Naleng Lake 4200 31°06 , 99°45 pollen Kramer et al. (2010)20 Yidun Lake 4470 30°18 , 99°33 pollen Shen et al. (2008b)21 Shayema Lake 2400 28°35 , 102°13 pollen Jarvis (1993)

Li et al. 943

The results show that 86.5% of the samples are correctly clas-

sified (Table 2). The samples from the alpine steppe belt are most

distinct from those from meadows according to discriminant

function 1, which accounts for 91.7% of variance (Figure 3).

However, a few samples of alpine meadow (S21, S23, S35) and

marsh meadow (S29, S37) are misclassified between these two

meadow types (Figures 2 and 3). This could reflect the similarities

in both vegetation compositions and pollen assemblages between

alpine and marsh meadows, since they share most taxa under a

predominance of Cyperaceae.

In summary, three major vegetation types are well distin-

guished by discriminant analysis. High values of probability of

modern analogue (mean 0.9) for modern pollen samples indicate

that major vegetation types are fairly reflected by their pollen

assemblages. All three major vegetation types are taken into con-

sideration in conducting discriminant analysis on fossil pollen

data of core NMLC-1.

Fossil pollen assemblages of core NMLC-1

A total of 75 pollen types are identified in the fossil pollen record

of NMLC-1 core from Nam Co. Fossil pollen assemblages are

dominated by Artemisia, Cyperaceae, and Poaceae. Three pollen

types account for 72% to 88% of the total pollen count for each

sample. Other herb pollen types with low percentages include

Chenopodiaceae, Thalictrum, Aster-type, Ranunculaceae, Caryo-

phyllaceae, Polygonum, Saussurea-type, Taraxacum-type, Genti-ana, Crassulaceae, and Cruciferae, etc. Tree and shrub

pollen types (6.2%) comprise Pinus, Betula, Alnus, Abies, Picea,

Ephedra, etc. Three pollen zones can be divided based on the

percentage variations of major pollen types from core NMLC-1

(Figure 4).

Zone NC1 (332–313 cm, 8400–7200 yr BP). This zone is char-

acterized by a predominance of Artemisia pollen (48.8%–64.6%,

0 1

Pro

babi

lity

of

mod

ern

anal

ogue

Sam

ple

code

Tree & Shrub Herb

0 10000 20 40 2 2 2 2 20 20 20 20 20 1 5 20 40 60 20 40 60 80 0.01 1 30 4750 4800%grains/g m

S5S6S7S8S9

S10S11S12S13S14S15S16S17S18

Alp

ine

step

pe

S2S3S4

S22S24S25S26S27S28S29S30S31S32S33S36S37

Mar

sh m

eado

w

2

S1S19S20S21S23S34S35

Alp

ine

mea

dow

Veg

etat

ion

1 2 3

Alp

ine

step

peA

lpin

e m

eado

wM

arsh

mea

dow

Vegetationzonal index

Figure 2. Pollen percentage diagram of topsoil samples from Nam Co area, with altitudes of sampling sites and results of discriminant analysis

Table 2. Discriminant analysis results of topsoil samples from Nam Co catchment in central Tibetan Plateau

Predicted group membership

Actual group No. of samples Alpine steppe Alpine meadow Marsh meadow

Alpine steppe 14 14a (100)b 0 0Alpine meadow 7 0 4 (57.1) 3 (42.9)Marsh meadow 16 0 2 (12.5) 14 (87.5)

Total number of samples: 37Number of misclassified samples: 5Percentage of samples correctly classified: 86.5%

aNumber of samples classified as that group.bPercentage of samples classified into that group.

944 The Holocene 21(6)

mean 57.3%) and high pollen concentrations (1.4×105grains/g).

Cyperaceae show an increase trend from the bottom to the top of

the zone (from 15.1% to 26.2%, mean 21.2%). Poaceae (5.0%),

Thalictrum (3.4%), Aster (1.8%), Ranunculaceae (1.2%) and

Polygonum (1.1%) are common. Percentages of tree and shrub

types (3.7%) and dryness indicators (1.3%) exhibit low values. A/

Cy ratios gradually decline from 4.0 to 1.8. Artemisia percentages

dramatically decrease from 60.8% to 53.3% at the upper part of the

zone (7600–7200 yr BP), where Cyperaceae increases.

Zone NC2 (313–270 cm, 7200–2800 yr BP). Percentages of

Artemisia consistently decline from 59.9% to 30.6%, whereas an

increase occurs in Cyperaceae pollen (from 16.9% to 42.7%) and

Poaceae pollen (from 1.7% to 9.1%). There is also an increase for

Ranunculaceae, Polygonum, Caryophyllaceae, tree and shrub

types within this zone. A/Cy ratios continually decline (from 3.4

to 0.7). Pollen concentration reached its highest value (2.6×105

grains/g) throughout the core around 5700 yr BP. The lowest

value of dryness indicators occurred around 7000–6000 yr BP,

and then it sharply increased during 6000 and 5500 yr BP. In a

period of 3200–2800 yr BP, Cyperaceae pollen markedly

increased to 42.1% at the expense of Artemisia pollen, associated

with declines of pollen concentrations and A/Cy ratios, and an

increase of dryness indicators.

Zone NC3 (270–5 cm, 2800–40 yr BP). There are increases in

the percentages of Saussurea-type, Taraxacum-type, Gentiana,

Crassulaceae, Androsace, Caryophyllaceae, Saxifragaceae, dry-

ness indicators of Chenopodiaceae and Cruciferae, and gymno-

sperm pollen types such as Pinus, Picea and Abies. Pollen

concentration sharply declines to 8.7×104 grains/g. This zone is

divided into two subzones. Subzone NC3-1 (270–214 cm, 2800–

1800 yr BP) is distinguished from NC3-2 (214–5 cm, 1800–40 yr

BP) by higher values of Artemisia (47.5%) and A/Cy ratio (1.8),

and lower values of Cyperaceae (28.0%) and Poaceae (5.0%). In

subzone NC3-2, percentages of Cyperaceae markedly increase to

36.0% with a decrease of Artemisia (39.6%), and A/Cy ratios

fluctuate around the value of 1.0. In the periods of 1700–1500 yr

BP and 800–150 yr BP, Cyperaceae apparently increased along

with a decrease of Artemisia. Cyperaceae pollen decreased, and

Artemisia increased together with dryness indicators at 1100–800

yr BP.

DiscussionModern pollen indications of alpine vegetation and climatePollen analysis on topsoil samples from the Nam Co area shows

that alpine vegetation types from altitudinal vegetation belts on

northern slope of Nyainqentanglha Mountains can be distin-

guished by their modern pollen assemblages. Pollen assemblages

are characterized by high percentages of Artemisia (25.1%) and

Poaceae (11.5%) for samples from alpine steppe with an average

elevation of 4745 m (range from 4732 to 4757 m). Alpine

meadow samples with an average elevation of 4800 m (range

from 4785 to 4819 m) and marsh meadow samples (mean eleva-

tion of 4763 m, range from 4735 to 4781 m) are distinguished by

predominating Cyperaceae pollen (>60%) (Figure 2). Results of

discriminant analysis further demonstrate that most of the mod-

ern pollen samples are typical for major vegetation types along

an elevation gradient, as suggested by high values of probability

of modern analogue (Figures 2 and 3). Thus, the discriminant

functions based on modern pollen assemblages from the Nam Co

area can be used to identify modern analogs for fossil pollen

assemblages.

Artemisia and Cyperaceae are the two major pollen types in

the Nam Co area as well as the central TP (Shen et al., 2008a;

Yu et al., 2001). Pollen ratio of A/Cy has been used as a semi-

quantitative measure for summer temperature changes in the cen-

tral and eastern Plateau, and as a useful indicator to distinguish

meadow and steppe vegetation (Herzschuh et al., 2006a). Previ-

ous studies showed that Cyperaceae dominates alpine meadow

(Kobresia) and alpine steppe (Carex), and thus modern pollen

assemblages in cold high-alpine environments of the east and

central TP, while high frequency of Artemisia occurs in both veg-

etation and modern pollen assemblages of subalpine and alpine

(shrub) steppe in the south and temperate steppe in the northeast

under warmer conditions (Herzschuh et al., 2006a; Yu et al.,

2001). Herzschuh (2007) further validated the applicability of A/

Cy ratio in reconstructions of temperature and vegetation in

alpine meadow and steppe regions using a set of lake surface sedi-

ments from the eastern and central TP.

Owing to low temperature and relatively high humidity on

the upper slopes, alpine meadow of K. pygmaea occurs between

the alpine steppe belt below and the alpine sparse vegetation

belt above in the central and western TP (Miehe and Miehe,

2000). In the Nam Co area, the alpine steppe belt of Artemisiaand Poaceae with abundant Artemisia pollen and high A/Cy

ratio (>1.0) is located on the lakeshore below 4800 m, while the

alpine meadow belt of Kobresia with dominant Cyperaceae

pollen and low A/Cy ratio (<1.0) occupies the upper slopes

between 4800 m and 5200 m where temperature is lower

according to the lapse rate, and higher effective moisture due to

the lower degree of evaporation than in the alpine steppe belt.

Therefore, the relative percentage changes of Artemisia and

Cyperaceae pollen (i.e. A/Cy ratio) in lake sediments can be

used to recover the shift of altitudinal vegetation belts which

are considered sensitive to temperature changes in alpine

mountain regions (Herzschuh et al., 2006b; Weng et al., 2007).

If the temperature declines, altitudinal vegetation belts would

Alpine steppe Marsh meadowAlpine meadow

Group centroid

31

2

- 5 0 5

4

2

0

-2

-4

Misclassified sample

Function 1

Func

tion

2

S5S8

S17S16

S13

S9S7

S12

S14

S11

S15

S18

S6S10

S19S1

S20 S34

S37S29

S25

S33 S32

S36S26

S4

S2S28

S27S24

S31

S3S21

S23 S30

S22S35

Figure 3. Ordination of 37 topsoil samples along discriminant functions 1 and 2

Li et al. 945

Her

b

0 1 2 3

1000

2000

3000

4000

5000

6000

7000

80000

Age (cal.yr BP)Depth (m)

Alpine steppe

Alpine meadow

Marsh meadow

Probability ofmodern analogue

Pal

eove

geta

tion Pin

usAb

ies& P

icea

Betul

a Alnus Fabac

eae

Ephe

dra Artem

isia

Poace

ae Cypera

ceae

Cheno

podia

ceae

Aster

-type

Ranun

culac

eae

Thalictr

um Polyg

onum

Saussu

rea-ty

pe

Tarax

acum

-type

Genti

ana

Crassu

lacea

e

Crucife

rae Andro

sace

Caryop

hylla

ceae

Pollen

conc

entra

tion

A/Cy r

atio

Saxifra

gace

ae

Tree

& S

hrub

Drynes

s ind

icator

s

grai

ns/g

%V

eget

atio

nzo

nal i

ndex

NC

2

NC

1

NC

3-2

NC

3-1

00.

51

010

35

55

320

4060

1010

3050

35

510

55

53

35

25

120

000

43

21

0.6

15

12

3w

et

dry

cold

war

m0

00

Zone

Fig

ure

4.Po

llen

spec

tra

of c

ore

NM

LC-1

and

pal

eove

geta

tion

reco

nstr

uctio

n ba

sed

on d

iscr

imin

ant

anal

ysis

. Dot

ted

fram

es m

ark

the

maj

or c

old

peri

ods.

Cur

ves

of r

econ

stru

cted

pal

eove

geta

tion,

pol

len

conc

entr

atio

n, d

ryne

ss in

dica

tors

and

A/C

y ra

tio a

re e

xhib

ited

in lo

wes

s sm

ooth

ing

lines

(sp

an 0

.04)

. Bla

ck b

ars

betw

een

the

axes

of a

ge a

nd d

epth

sho

w d

atin

g po

ints

946 The Holocene 21(6)

shift downward to the lakeshore. Consequently, the alpine

steppe belt around the lake shrinks with a decline of its upper

altitude limit, and the alpine meadow belt distributes more

close to the lake basin and inputs more Cyperaceae pollen,

resulting in lower A/Cy ratios in the sediments. On the contrary,

altitudinal vegetation belts move upwards with increasing tem-

perature. Alpine steppe would expand to upper slopes, and

alpine meadow belt and alpine sparse vegetation belt above

would shift to upper slopes far away from the lake basin, caus-

ing a reduction of Cyperaceae pollen but an increase of Arte-misia pollen in lake sediments which lead to higher A/Cy ratios.

In the interpretation of NMLC-1 pollen assemblages, we use

the A/Cy ratio to indicate temperature change and vertical shift

of alpine vegetation belts.

The pollen ratio of Artemisia to Chenopodiaceae (A/C) is

widely used as an indicator for effective moisture and desert/

steppe transition in the Middle East (El-Moslimany, 1990), cen-

tral arid Asia (Sun et al., 1994), and the western and northeastern

TP (Cour et al., 1999; Gasse et al., 1991; Liu et al., 1998; Van

Campo et al., 1996; Zhao and Herzschuh, 2009). A higher A/C

value points to more steppe-like vegetation and wetter conditions,

while a lower A/C value indicates more desert-like vegetation and

drier conditions (Herzschuh, 2007). However, there is such few

Chenopodiaceae and abundant Artemisia in modern and fossil

pollen records in the steppe and meadow regions of the TP that

A/C ratios are commonly high (>50) with large statistical errors.

Therefore, we use the pollen abundance of dryness indicators

instead of the A/C ratio to reveal drought periods and moisture

changes in the Nam Co area.

Modern pollen assemblages from topsoil samples in the Nam

Co area are dominated by herb pollen such as Artemisia, Cypera-

ceae and Poaceae. The results of discriminant analysis suggest

that modern pollen assemblages are consistent with local vegeta-

tion. Lu et al. (2010) analysed airborne pollen assemblages in the

Nam Co area, and found that they were dominated by local herba-

ceous taxa (87%) including Cyperaceae and Artemisia. Peak con-

centrations of these pollen types occurred in the flowering seasons

of spring, summer, and early autumn. Therefore, these results

suggest that vegetation within the catchment area of Nam Co is

the dominant pollen source for this lake.

Arboreal pollen types take up a very small proportion in mod-

ern pollen assemblages of the Nam Co area. Nowadays there is no

forest in this region. It has been reported that arboreal pollen types

such as Pinus could be transported over 400 km to the central and

western TP by atmospheric circulation (Wu and Xiao, 1995). Lu

(2008) demonstrated that Pinaceae pollen could be transported to

the Nam Co area over long distances from conifer forests located

in the southern and southwestern regions, and that Betula pollen

could come from birches growing in the southern and northeast-

ern Plateau. Therefore, arboreal pollen types with very low per-

centages in both modern and fossil pollen assemblages of Nam

Co area are considered as exotic types from far-distant vegetation

sources.

Inferred vertical shifts of alpine vegetation belts and climate changes

Fossil pollen assemblages of core NMLC-1 and results of dis-

criminant analysis reveal a detailed history of vertical shifts of

altitudinal vegetation belts and climate changes in the Nam Co

area since 8400 yr BP (Figure 4).

During the period 8400–7200 yr BP, a high probability of

modern analogue (0.96) and stable values of vegetation zonal

indices (1.0) indicate that alpine steppe prevailed over the Nam

Co area. The typical steppe element of Artemisia (57.3%) domi-

nated pollen assemblages. Its percentages markedly exceed the

average value of topsoil samples from alpine steppe (25.1%).

Cyperaceae pollen occurred in the lowest percentage of the core.

Pollen assemblages suggest that Artemisia-dominated steppe

had much greater distribution range and higher upper altitude

limit, while alpine meadow belt and alpine sparse vegetation

belt occurred only on slopes at elevations higher than that of

today with very limited vertical extents. High values of A/Cy

ratio and relatively low percentages of dryness indicators sug-

gest that the climate in the Nam Co area was characterized by

high temperature and moderate moisture during this period. This

warm and moderate humid condition is also supported by the

increases of TOC and n-alkanes of benthonic and terraneous

plants (Zhu et al., 2008), and ostracod assemblages (Xie et al.,

2009) recorded in the same NMLC-1 core. A cold and slight

drought interval around 7600–7200 yr BP is indicated by a

decline of A/Cy ratio, an increase of dryness indicators, and a

peak value of CaCO3 content (Zhu et al., 2008). This cold/dry

interval seems a centennial event in the central and western TP.

Rapid decrease of precipitation between 7500 and 7000 yr BP is

found in a geochemical record from Ahung Co (Morrill et al.,

2006). Fossil pollen assemblages from Lake Zigetang with decreases

of pollen concentration and A/Cy ratio around 7.4 kyr BP indi-

cate the regression of Artemisia-dominated steppe (Herzschuh

et al., 2006a). Paleoenvironmental records from Bangong Co in

the western Plateau show a drought around 8–7.7 kyr BP (Van

Campo et al., 1996).

In the mid to late Holocene (7200–2800 yr BP), results of

discriminant analysis show that vegetation around the lake was

still dominated by alpine steppe. However, Artemisia gradually

decreased to 30.6%, whereas Cyperaceae, Ranunculaceae,

Caryophyllaceae and Saussurea-type increased. The variations

in pollen percentages indicate that the alpine steppe belt gradu-

ally shrank, while the alpine meadow belt and the alpine sparse

vegetation belt on the upper slopes extended downwards towards

the lake, associated with lowering of their lower altitude limits.

Temperature in the Nam Co area continually declined, and mois-

ture was highest during 7000–6000 yr BP and then sharply

dropped, as reflected by decreasing A/Cy ratios and fluctuations

of dryness indicators. A severe cold/dry event occurring at 3200–

2800 yr BP led to alpine meadow thriving in the Nam Co area, as

indicated by vegetation zonal indices of ~1.6, low A/Cy ratios

around 0.8, and relatively high percentages of dryness indicators

(~2%). This event, characterized by downward expansion of

alpine meadow belt in our pollen record, is consistent with west-

ward movement of the meadow-steppe ecotone and an establish-

ment of typical alpine meadow around 2800 yr BP in the Co

Ngion area northeast of Nam Co under precipitation near present

values and cold conditions (Shen et al., 2008a).

During the late Holocene (2800 yr BP to the present), vegeta-

tion and climate were characterized by their instability. The

marked increases in typical components of alpine sparse vegeta-

tion such as Saussurea-type, Gentiana, Crassulaceae, Androsace,

Caryophyllaceae and Saxifragaceae since 2800 yr BP indicate

that the alpine sparse vegetation belt markedly extended during

the late Holocene. In the first stage, between 2800 and 1800 yr

BP, an average vegetation zonal index of 1.1 indicated that alpine

Li et al. 947

steppe recovered with a probable ascension of its upper limit,

while the alpine meadow belt withdrew towards the upper slopes,

under relatively warm conditions as indicated by A/Cy ratios of

1.8. In the last stage from 1800 yr BP to the present, vegetation

zonal index gradually increased with frequent fluctuations around

value of 1.5 (boundary between alpine steppe and alpine meadow),

suggesting an extension of alpine meadow belt and instability of

altitudinal vegetation belts under climate cooling, indicated by

reduction of A/Cy ratios. Furthermore, variations of A/Cy ratios

and percentages of dryness indicators revealed two cold/wet

intervals at 1700–1500 yr BP and 800–150 yr BP (‘Little Ice

Age’, LIA) and a relatively warm/dry interval at 1100–800 yr BP

(‘Medieval Warm Period’, MWP) (Mügler et al., 2010; Zhang

et al., 2002; Zhu et al., 2008).

Regional comparison of Holocene climate in the TP

Holocene climate in the Nam Co area is characterized by warm/

wet conditions in the early to mid Holocene, and frequent fluctua-

tions during the late Holocene as inferred from pollen assem-

blages of core NMLC-1.

A decreasing trend of A/Cy ratios and a general downward

shift of altitudinal vegetation belts on northern slope of Nyain-

qentanglha Mountains indicate a climatic cooling throughout the

last 8400 years (Figure 4). It seems to agree with reconstructed

climates in other regions of the TP. Fossil pollen assemblages of

Lake Zigetang suggest a general cooling trend in the Holocene

and zonal vegetation transition from Artemisia-dominated steppe

to Cyperaceae-dominated alpine steppe at 4.4 kyr BP (Herzschuh

et al., 2006a). The pollen record of Lake Koucha in the northeast-

ern TP indicates a temperature decline since 8–7 kyr BP and a

vegetation transition from warm/semi-wet temperate steppe to

cold/wet alpine meadow around 6.6 kyr BP (Herzschuh et al.,

2009b). This cooling trend of Holocene climate is correlated well

with variations of summer solar insolation at 30°N (Berger and

Loutre, 1991), reflecting a consistent influence of insolation

forcing on vegetation and climate changes in high-alpine envi-

ronments of the central TP (Herzschuh et al., 2006a).

The moisture in the Nam Co area culminated at around 7–6

kyr BP, as indicated by low percentages of dryness indicators and

high pollen concentrations in the pollen record of NMLC-1 core,

and by low contents of CaCO3 due to an increasing lake water

volume (Zhu et al., 2008). This maximum moisture in the mid

Holocene widely recorded in the central TP (Herzschuh et al.,

2006a; Wu et al., 2006) could be attributed to an integrated func-

tion of many factors, including lower degree of evaporation due

to temperature decline since the early Holocene (Herzschuh et al.,

2009b; Wu et al., 2006), soil moisture supply from melting ice

and snow (Kashiwaya et al., 1995) on the Nyainqentanglha

Mountains, and above all, an enhanced monsoon precipitation

which was caused by the intensification of Southwest Indian

Monsoon lasting from 9.2 to 6 kyr BP, lagging the insolation max-

imum of the Northern Hemisphere by about 2000 to 3000 years

because of the control from glacial boundary conditions in high

northern latitudes (Fleitmann et al., 2003; Overpeck et al., 1996).

Warm and wet climate during the early and mid Holocene is

widely suggested by numerous paleoenvironment records with a

temporal succession across the Plateau, e.g. (1) western TP: pol-

len and diatom records from Sumxi Co (11.4–6.5 kyr BP, Gasse

et al., 1991; Van Campo and Gasse, 1993) and Bangong Co

(9.5–6.2 kyr BP, Van Campo et al., 1996), oxygen isotope record

of Guliya ice core (warm condition during 11.5–7.2 kyr BP,

Thompson et al., 1997); (2) central TP: paleoclimate records

from Selin Co (9.6–6 kyr BP, Sun et al., 1993; maximum mon-

soon precipitation during 8.4–5.5 kyr BP, Gu et al., 1994),

NMLC-1 pollen record (warm/wet climate during 8.4–6 kyr BP)

and higher lake level of Nam Co during 9.6–5.2 kyr BP (Schütt

et al., 2010), sedimentary record from Lake Cuoe (10–5.7 kyr

BP, Wu et al., 2006), Lake Zigetang pollen record (10.8–4.4 kyr

BP, Herzschuh et al., 2006a) and geochemistrical evidence (9–

5.6 kyr BP, Wu et al., 2007); (3) southern TP: pollen records

from Hidden Lake and Ren Co (about 8–5 kyr BP, Tang et al.,

1999); multiproxy records from Lake Chen Co (warm/wet con-

dition at 9–6 kyr BP, Zhu et al., 2009); (4) northern and north-

eastern TP: pollen records from Dunde ice core (10–4.8 kyr BP,

Liu et al., 1998), Lake Qinghai (around 10.2–4 kyr BP, Her-

zschuh et al., 2010a; Shen et al., 2005), and Lake Luanhaizi (9–6

kyr BP, Herzschuh et al., 2010b); (5) eastern TP: multiproxy

records from Ximencuo (10–4.4 kyr BP, Zhang and Mischke,

2009), carbon isotope record from Hongyuan peat (10.8–5.5 kyr

BP, Hong et al., 2003), pollen records from Lake Naleng (10–4.4

kyr BP, Kramer et al., 2010), Yidun Lake (9.8–2.5 kyr BP, Shen

et al., 2006), and Lake Shayema (10.9–2.3 kyr BP, Jarvis, 1993;

Shen, 2003). In summary, from recent studies mentioned above,

the termination of maximum monsoon precipitation or moisture

condition occurred at around 6.5–6.2 kyr BP in the western TP,

6–4.4 kyr BP in the central and southern Plateau, about 6–4 kyr

BP in the northern and northeastern regions, and around 4.4–2.3

kyr BP in the eastern margins (Figure 5). This temporal succes-

sion from the northwest to the southeast TP illustrates a hypoth-

esized southeastward retreat of Southwest Monsoon front along

a modern precipitation gradient during the mid to late Holocene

(as reviewed by Liu et al., 1998; Shen, 2003; Shen and Tang,

1996; Shen et al., 2008b), although dating uncertainty exists

among these records. Some exceptions with maximum precipi-

tation in the late Holocene were presented in the northeastern TP

(e.g. Lake Koucha) and attributed to regional effects such as a

decline of evaporation–precipitation ratio in association with

decreasing temperatures after the early Holocene (Herzschuh

et al., 2009b; Kramer et al., 2010).

Frequent fluctuations of late-Holocene climate under cooler

and drier conditions, reflected by pollen record and other prox-

ies (e.g. contents of TOC and CaCO3, Zhu et al., 2008) in Nam

Co, can be also deduced from other sites of the TP, e.g. Lake

Qinghai (Shen et al., 2005), Hongyuan peat (Hong et al., 2003),

Lake Kusai (Liu et al., 2009), Lake Cuoe (Wu et al., 2006),

Lake Zigetang (Herzschuh et al., 2006a) and Co Ngion (Shen

et al., 2008a). Despite dating uncertainty, three centennial-scale

cooling intervals in the Nam Co area around 0.8–0.15 kyr BP,

1.7–1.1 kyr BP and 3.2–2.8 kyr BP can be chronologically cor-

related to cold intervals documented in the oxygen isotope

record of Guliya ice core (Thompson et al., 1997), and to LIA

and ice-rafted debris events in the North Atlantic that culmi-

nated around 0.4 kyr BP, 1.4 kyr BP and 2.8 kyr BP (Bond et al.,

2001), while two of the relatively warm periods coincide with

the in-between warm intervals in the North Atlantic (Figure 6).

A warm/dry interval during 1100–800 yr BP in the Nam Co area

could correspond to the MWP (Liu et al., 1998; Zhang, et al.,

2002), and the latest cold/wet interval of 800–150 yr BP to the

LIA recorded widely in the TP (Liu et al., 1998, 2009; Shen

et al., 2008a; Zhang et al., 2002; Zhu et al., 2008). Correlation

of late-Holocene climate between the central, western TP and

948 The Holocene 21(6)

the North Atlantic implies that influences from abrupt climatic

events in the high northern latitudes reached the central TP via

the Westerlies after the marked weakening of both insolation

and Southwest Monsoon during the mid to late Holocene

(Overpeck et al., 1996; Shen et al., 2008b). Therefore, the

modern climate in the central TP, characterized by alternate

influences from the Westerlies and the Southwest Monsoon

(Lin et al., 2008; You et al., 2007; Zhu et al., 2008), was possibly

-12

0 1000 2000 3000 4000 0

4

8

MWP

Gorth

Atla

ntic

mat

ite-s

tain

edgr

ains

(%)

(a)1

Age (cal. yr BP)

MLC

-1C

y ra

tio

1.5

2

12

-18

-16

-1412 0

Guliya ice core

18O (‰

)N

oH

emg

(b)

(c)

2

warm

NM

LC-1

TOC

(%)

NM

A/C

0.51

(d)1

1.5

0.510

2C-1 dica

tors

)

cold

t

0 1000 2000 3000 4000

NM

LC-1

CaC

O3

(%)

(e)32

4

NM

LCD

ryne

ss in

d(%

)

(f)

20304050

Age (cal yr BP)

wet

dry

Age (cal. yr BP)

LIA

Figure 6. Comparisons among paleoclimate records in Nam Co, ice-core oxygen isotope records in the western TP, and drift ice records in high northern latitudes in the late Holocene. (a) Stack of drift ice records from MC52-VM29191 and MC21-GGC22 cores in the North Atlantic (Bond et al., 2001). Numbers of 0, 1 and 2 indicate LIA and other two ice-rafted debris events, respectively. (b) 18O record of Guliya ice core (Thompson et al., 1997); (c) and (e) are 5-point smoothed curves of A/Cy ratio and dryness indicators from NMLC-1 pollen records; (d) and (f) show the variations of contents of TOC and CaCO3 from NMLC-1 core (Zhu et al., 2008). Dotted frames highlight cold intervals, while grey bars indicate warm intervals

2000

3000

4000

5000

6000

7000

8000

9000

10000

11000

12000

2000

3000

4000

5000

6000

7000

8000

9000

10000

11000

12000

Age

(yr B

P)

west central and south north and northeast east

Age (yr B

P)

Figure 5. Summary of inferred SWM records across the TP during the early to mid Holocene. Dark and light grey frames represent wet and dry periods, respectively. Dotted horizontal lines highlight the time span for the termination of maximum precipitation or moisture in various regions. References are given in the text and Table 1

Li et al. 949

established within the late Holocene (Shen and Tang, 1996;

Tang et al., 2009; Wu et al., 2006).

ConclusionsModern pollen rains around Nam Co are investigated in this

study. The results are then applied to interpret fossil pollen assem-

blages of core NMLC-1 from Nam Co for recovering the history

of vegetation and climate changes in the central TP since 8400 yr

BP. The main conclusions are summarized below.

(1) Pollen analysis on 37 topsoil samples from altitudinal vege-

tation belts on the northern slope of Nyainqentanglha Moun-

tains shows that alpine steppe and alpine meadow along an

elevation gradient can be distinguished by their modern pol-

len assemblages. Samples from alpine steppe contain high

proportions of Artemisia and Poaceae, while those from

alpine meadow on upper slopes and patchy marsh meadow

are characterized by a predominance of Cyperaceae. Results

of discriminant analysis correctly classify 86.5% of topsoil

samples in terms of vegetation types.

(2) The NMLC-1 pollen record reveals a general downward

shift of altitudinal vegetation belts since 8.4 kyr BP, and a

major extension of alpine meadow belt and alpine sparse

vegetation belt towards the lakeshore in the late Holocene.

(3) Pollen-inferred Holocene climate in the Nam Co area is

characterized by a climatic cooling throughout the last 8400

years following the decline of solar insolation, with warmer

and wetter climate in the early to mid Holocene resulting

from an intensified Southwest Monsoon and strong insola-

tion, and frequent climate fluctuations in the late Holocene,

which can be linked to the high northern latitudes on a cen-

tennial scale.

(4) The termination timing of maximum moisture in the Nam

Co area around 6000–5500 yr BP provides a link between

the northwest and southeast TP for an interpretation of

southeastward movement of Southwest Monsoon during the

mid to late Holocene.

AcknowledgementsWe would like to thank two anonymous reviewers and the editors

for their constructive comments and suggestions. We are grate-

ful to Xiao Lin, Jianting Jü and Yong Wang for field assistance,

and Luo Wang, Caiming Shen and Patrick Rioual for helpful

discussions and language corrections. This work was supported

by Chinese Academy of Sciences (grant number KZCX2-YW-

Q09-03-02), the National Basic Research Programme of China

(grant number 2005CB422002) and the National Science Founda-

tion of China (grant numbers 40771216, 41071131, 40701194).

ReferencesBerger A and Loutre MF (1991) Insolation values for the climate of the last 10

million years. Quaternary Science Reviews 10: 297–317.

Bond G, Kromer B, Beer J, Muscheler R, Evans MN, Showers W et al. (2001)

Persistent solar influence on North Atlantic climate during the Holocene.

Science 294: 2130–2136.

Chen JH, Yang YP, Lu XM, Li ZQ, Zhang M and Xu J (2011) Primary study of

plant community in the Nam Co Basin. In: Kang SC (ed.) Modern Envi-

ronmental Processes and Changes in the Nam Co Basin, Tibetan Plateau.

Beijing: Meteorological Press (in press).

Cour P, Zheng Z, Duzer D, Calleja M and Yao Z (1999) Vegetational and cli-

matic significance of modern pollen rain in northwestern Tibet. Review of Palaeobotany and Palynology 104: 183–204.

El-Moslimany AP (1990) Ecological significance of common nonarboreal pol-

len: Examples from drylands of the Middle East. Review of Palaeobotany and Palynology 64: 343–350.

Fægri K and Iversen J (1989) Textbook of Pollen Analysis (4th Edition). Lon-

don: John Wiley and Sons.

Fleitmann D, Burns SJ, Mudelsee M, Neff U, Kramers J, Mangini A et al.

(2003) Holocene forcing of the Indian monsoon recorded in a stalagmite

from southern Oman. Science 300: 1737–1739.

Gasse F, Arnold M, Fontes JC, Fort M, Gibert E, Huc A et al. (1991) A 13,000-

year climate record from western Tibet. Nature 353: 742–745.

Gu ZY, Liu JQ, Yuan BY, Liu DS and Zhang GY (1994) Lacustrine authigenic

deposition expressive of environment and the sediment record from Siling

Co, Xizang (Tibet), China. Quaternary Sciences 2: 162–174 (in Chinese).

Herrmann M, Lu XM, Berking J, Schütt B, Yao TD and Mosbrugger V (2010)

Reconstructing Holocene vegetation and climate history of Nam Co area

(Tibet), using pollen and other palynomorphs. Quaternary International 218: 45–57.

Herzschuh U (2007) Reliability of pollen ratios for environmental reconstruc-

tions on the Tibetan Plateau. Journal of Biogeography 34: 1265–1273.

Herzschuh U, Birks HJB, Liu XQ, Kubatzki C and Lohmann G (2010a) What

caused the mid-Holocene forest decline on the eastern Tibet-Qinghai Pla-

teau? Global Ecology and Biogeography 19: 278–286.

Herzschuh U, Birks HJB, Mischke S, Zhang CJ and Böhner J (2010b)

A modern pollen–climate calibration set based on lake sediments from

the Tibetan Plateau and its application to a Late Quaternary pollen record

from the Qilian Mountains. Journal of Biogeography 37: 752–766.

Herzschuh U, Birks HJB, Ni J, Zhao Y, Liu HY, Liu XQ et al. (2009a) Holocene

land-cover changes on the Tibetan Plateau. The Holocene 20: 91–104.

Herzschuh U, Kramer A, Mischke S and Zhang CJ (2009b) Quantitative

climate and vegetation trends since the late glacial on the northeastern

Tibetan Plateau deduced from Koucha Lake pollen spectra. QuaternaryResearch 71: 162–171.

Herzschuh U, Kürschner H and Mischke S (2006b) Temperature variability

and vertical vegetation belt shifts during the last 50000 yr in the Qil-

ian Mountains (NE margin of the Tibetan Plateau, China). QuaternaryResearch 66: 133–146.

Herzschuh U, Winter K, Wünnemann B and Li S (2006a) A general cooling trend

on the central Tibetan Plateau throughout the Holocene recorded by the

Lake Zigetang pollen spectra. Quaternary International 154–155: 113–121.

Hong YT, Hong B, Lin QH, Zhu YX, Shibata Y, Hirota M et al. (2003) Correla-

tion between Indian Ocean summer monsoon and North Atlantic climate

during the Holocene. Earth and Planetary Science Letters 211: 371–380.

Institute of Geography (1990) Map of the Qinghai-Tibet Plateau. Beijing:

Science Press (in Chinese).

Jarvis DI (1993) Pollen evidence of changing Holocene monsoon climate in

Sichuan Province, China. Quaternary Research 39: 325–337.

Kashiwaya K, Masuzawa T, Morinaga H, Yaskawa K, Yuan BY, Liu JQ et al.

(1995) Changes in hydrological conditions in the central Qing-Zang

(Tibetan) Plateau inferred from lake bottom sediments. Earth and Plan-etary Science Letters 135: 31–39.

Kramer A, Herzschuh U, Mischke S and Zhang CJ (2010) Holocene treeline

shifts and monsoon variability in the Hengduan Mountains (southeastern

Tibetan Plateau), implications from palynological investigations. Palaeo-geography, Palaeoclimatology, Palaeoecology 286: 23–41.

Li MH, Kang SC, Zhu LP, You QL, Zhang QG and Wang JB (2008) Mineral-

ogy and geochemistry of the Holocene lacustrine sediments in Nam Co,

Tibet. Quaternary International 187: 105–116.

Li YC, Xu QH, Zhang LY, Wang XL, Cao XY and Yang XL (2009) Modern

pollen assemblages of the forest communities and their relationships with

vegetation and climate in northern China. Journal of Geographical Sci-ences 19: 643–659.

Lin X, Zhu LP, Wang Y, Wang JB, Xie MP, Ju JT et al. (2008) Environmental

changes reflected by n-alkanes of lake core in Nam Co on the Tibetan

Plateau since 8.4 ka BP. Chinese Science Bulletin 53: 3051–3057.

Liu KB and Lam NS-N (1985) Paleovegetational reconstruction based on mod-

ern and fossil pollen data: An application of discriminant analysis. Annalsof the Association of American Geographers 75: 115–130.

Liu KB, Zhao ZJ and Thompson LG (1998) A pollen record of Holocene cli-

matic changes from the Dunde ice cap, Qinghai-Tibetan Plateau. Geology26: 135–138.

Liu XQ, Dong HL, Yang XD, Herzschuh U, Zhang EL, Stuut J-BW et al.

(2009) Late Holocene forcing of the Asian winter and summer monsoon

as evidenced by proxy records from the northern Qinghai–Tibetan Pla-

teau. Earth and Planetary Science Letters 280: 276–284.

Lu HY, Wang SM, Wu NQ, Tong GB, Yang XD, Shen CM et al. (2001) A new

pollen record of the last 2.8 Ma from the Co Ngoin, central Tibetan Pla-

teau. Science in China Series D: Earth Science 44: 292–300.

Lu XM (2008) Modern pollen distribution and its relationship with vegetation

and climate in Nam Co catchment, Tibet. Postdoctoral research report,

Institute of Geology and Geophysics, Beijing (in Chinese).

950 The Holocene 21(6)

Lu XM, Herrmann M, Mosbrugger V, Yao TD and Zhu LP (2010) Airborne

pollen in the Nam Co Basin and its implication for palaeoenvironmental

reconstruction. Review of Palaeobotany and Palynology 163: 104–112.

Lynch EA (1996) The ability of pollen from small lakes and ponds to sense

fine-scale vegetation patterns in the Central Rocky Mountains, USA.

Review of Palaeobotany and Palynology 94: 197–210.

Miehe G and Miehe S (2000) Environmental changes in the pastures of Xizang.

Marburger Geographische Schriften 135: 282–312.

Morrill C, Overpeck JT, Cole JE, Liu KB, Shen CM and Tang LY (2006) Holo-

cene variations in the Asian monsoon inferred from the geochemistry of

lake sediments in central Tibet. Quaternary Research 65: 232–243.

Mügler I, Gleixner G, Günther F, Mäusbacher R, Daut G, Schütt B et al. (2010)

A multi-proxy approach to reconstruct hydrological changes and Holo-

cene climate development of Nam Co, Central Tibet. Journal of Paleo-limnology 43: 625–648.

Overpeck JT, Anderson D, Trumbore S and Prell W (1996) The southwest Indian

Monsoon over the last 18,000 years. Climate Dynamics 12: 213–225.

Reimer PJ, Baillie MGL, Bard E, Bayliss A, Beck JW, Bertrand CJH et al.

(2004) Intcal04 terrestrial radiocarbon age calibration, 0–26 cal kyr BP.

Radiocarbon 46: 1029–1058.

Schütt B, Berking J, Frechen M, Frenzel P, Schwalb A and Wrozyna C (2010)

Late Quaternary transition from lacustrine to a fluvio-lacustrine environ-

ment in the north-western Nam Co, Tibetan Plateau, China. QuaternaryInternational 218: 104–117.

Shen CM (2003) Millennial-scale variations and centennial scale events in the

Southwest Asian monsoon: Pollen evidence from Tibet. Ph.D. disserta-

tion, Louisiana State University, Baton Rouge.

Shen CM and Tang LY (1996) Pollen evidence of changing Holocene mon-

soon on Qinghai-Xizang Plateau. Acta Micropalaeontologica Sinica 13:

433–436 (in Chinese).

Shen CM, Liu KB, Morrill C, Overpeck JT, Peng JL and Tang LY (2008a) Eco-

tone shift and major droughts during the mid-late Holocene in the central

Tibetan Plateau. Ecology 89: 1079–1088.

Shen CM, Liu KB, Tang LY and Overpeck JT (2006) Quantitative relationships

between modern pollen rain and climate in the Tibetan Plateau. Review of Palaeobotany and Palynology 140: 61–77.

Shen CM, Liu KB, Tang LY and Overpeck JT (2008b) Numerical analysis

of modern and fossil pollen data from the Tibetan Plateau. Annals of the Association of American Geographers 98: 755–772.

Shen CM, Tang LY, Wang SM, Li CH and Liu KB (2005) Pollen records and

time scale for the RM core of the Zoige Basin, northeastern Qinghai-

Tibetan Plateau. Chinese Science Bulletin 50: 553–562.

Shen J, Liu XQ, Wang SM and Matsumoto R (2005) Palaeoclimatic changes

in the Qinghai Lake area during the last 18000 years. Quaternary Inter-national 136: 131–140.

Song MH, Zhou CP and Ouyang H (2004) Distributions of dominant tree spe-

cies on the Tibetan Plateau under current and future climate scenarios.

Mountain Research and Development 24: 166–173.

SPSS Inc. (1999) SPSS Base 10.0 for Windows user’s guide. Chicago IL: SPSS Inc.

Sugden JM and Meadows ME (1989) The use of multiple discriminant analysis

in reconstructing recent vegetation on the Nuweveldberg, South Africa.

Review of Palaeobotany and Palynology 60: 131–147.

Sun XJ, Du NQ, Chen YS, Gu ZY, Liu JQ and Yuan BY (1993) Holocene

palynological records in Lake Selincuo, northern Xizang (Tibet). ActaBotanica Sinica 35: 943–950 (in Chinese).

Sun XJ, Du NQ, Weng CY, Lin RF and Wei KQ (1994) Paleovegetation and

paleoenvironment of Manasi Lake, Xinjiang, NW China during the last

14000 years. Quaternary Sciences 3: 239–248 (in Chinese).

Tang LY, Shen CM, Li CH, Peng JL, Liu H, Liu KB et al. (2009) Pollen inferred

vegetation and environment changes in central Tibetan Plateau since 8200

year BP. Science in China Series D: Earth Science 52: 1104–1114.

Tang LY, Shen CM, Liu KB and Overpeck JT (1999) New high-resolution

pollen records from two lakes in Xizang (Tibet). Acta Botanica Sinica 41:

896–902 (in Chinese).

Tang LY, Shen CM, Liu KB, Yu SY and Li CH (2004) Climatic changes in the

Southeastern Qinghai-Tibetan Plateau since the Last Glacial Maximum –

Pollen records from Southeastern Tibet. Science in China Series D: EarthScience 34: 436–442 (in Chinese).

Thompson LG, Yao TD, Davis ME, Henderson KA, Thompson EM, Lin PN

et al. (1997) Tropical climatic instability: The last glacial cycle from a

Qinghai-Tibetan ice core. Science 276: 1821–1825.

Tibetan Investigation Group (1988) Vegetation of Xizang (Tibet). Beijing: Sci-

ence Press (in Chinese).

Van Campo E and Gasse F (1993) Pollen- and diatom-inferred climatic and

hydrological changes in Sumxi Co Basin (Western Tibet) since 13000 yr

BP. Quaternary Research 39: 300–313.

Van Campo E, Cour P and Hang SX (1996) Holocene environmental changes

in Bangong Co basin (Western Tibet) Part 2: The pollen record. Palaeo-geography, Palaeoclimatology, Palaeoecology 120: 49–63.

Wang FX, Qian NF, Zhang YL and Yang HQ (1997) Pollen Flora of China.

Beijing: Science Press (in Chinese).

Wang JB, Zhu LP, Daut G, Ju JT, Lin X, Wang Y et al. (2009) Bathymetric

survey and modern limnological parameters of Nam Co, central Tibet.

Journal of Lake Sciences 21: 128–134 (in Chinese).

Weng CY, Hooghiemstra H and Duivenvoorden JF (2007) Response of pollen

diversity to the climate-driven altitudinal shift of vegetation in the Colom-

bian Andes. Philosophical Transactions of the Royal Society B: BiologicalSciences 362: 253–262.

Wu YH, Lücke A, Jin ZD, Wang SM, Schleser GH, Battarbee RW et al. (2006)

Holocene climate development on the central Tibetan Plateau: A sedimen-

tary record from Cuoe Lake. Palaeogeography, Palaeoclimatology, Pal-aeoecology 234: 328–340.

Wu YH, Lücke A, Wünnemann B, Li SJ and Wang SM (2007) Holocene

climate change in the Central Tibetan Plateau inferred by lacustrine

sediment geochemical records. Science in China Series D: Earth Sciences 50: 1548–1555.

Wu YS and Xiao JY (1995) A preliminary study on modern pollen rain of

Zabuye Salt Lake area, Xizang. Acta Botanica Yunnanica 17: 72–78 (in

Chinese).

Wu ZH, Zhao XT, Wu ZH, Wu XC, Zhou CJ, Yan FH et al. (2004) Geological

records of climate and environment changes during the Holocene in the

Nam Co lake and its adjacent areas. Journal of Glaciology and Geocryology 26: 275–283 (in Chinese).

Xi YZ and Ning JC (1994) Study on pollen morphology of plants from dry and

semidry area in China. Yushania 11: 119–191 (in Chinese).

Xie MP, Zhu LP, Peng P, Wang JB, Wang Y and Schwalb A (2009) Ostracod

assemblages and their environmental significance from the lake core of

the Nam Co on the Tibetan Plateau 8.4 ka BP. Journal of Geographical Sciences 4: 387–402.

You QL, Kang SC, Li CL, Li MS and Liu JS (2007) Variation features of mete-

orological elements at Namco Station, Tibetan Plateau. Meteorological Monthly 33: 54–60 (in Chinese).

Yu G, Tang LY, Yang XD, Ke XK and Harrison SP (2001) Modern pollen sam-

ples from alpine vegetation on the Tibetan Plateau. Global Ecology and Biogeography 10: 503–519.

Zhang CJ and Mischke S (2009) A Lateglacial and Holocene lake record from

the Nianbaoyeze Mountains and inferences of lake, glacier and climate

evolution on the eastern Tibetan Plateau. Quaternary Science Reviews 28:

1970–1983.

Zhang EL, Shen J, Wang SM, Xia WL and Jin ZD (2002) Climate and environ-

ment change during the past 900 years in Qinghai Lake. Journal of Lake Sciences 14: 32–38 (in Chinese).

Zhang JW, Wang JT, Chen WL and Li BS (1981) On latitudinal zoning of

vegetation in the Qinghai-Xizang (Tibet) Plateau. Scientia Sinica 24:

1154–1165.

Zhang SR (2004) Revision of Kobresia (Cyperaceae) in Xizang (Tibet). ActaPhytotaxonomica Sinica 42: 194–221.

Zhao XT, Zhu DG, Yan FH, Wu ZH, Ma ZB and Mai XS (2003) Climatic

change and lake-level variation of Nam Co, Xizang since the last intergla-

cial stage. Quaternary Sciences 23: 41–52 (in Chinese).

Zhao Y and Herzschuh U (2009) Modern pollen representation of source

vegetation in the Qaidam Basin and surrounding mountains, north-

eastern Tibetan Plateau. Vegetation History and Archaeobotany 18:

245–260.

Zheng D (1996) The system of physico-geographical regions of the Tibetan

Plateau. Science in China Series D: Earth Science 26: 336–341 (in

Chinese).

Zhu DG, Meng XG, Zhao XT, Shao ZG, Feng XY, Yang CB et al. (2004)

On the Quaternary Environmental Evolution of the Nam Co area, Tibet.Beijing: Science Press (in Chinese).

Zhu LP, Wu YH, Wang JB, Lin X, Ju JT, Xie MP et al. (2008) Environmental

changes since 8.4 ka reflected in the lacustrine core sediments from Nam

Co, central Tibetan Plateau, China. The Holocene 18: 831–839.

Zhu LP, Zhen XL, Wang JB, Lu HY, Xie MP, Kitagawa H et al. (2009)

A ~30,000-year record of environmental changes inferred from Lake

Chen Co, Southern Tibet. Journal of Paleolimnology 42: 343–358.